XGBoost server is an implementation for serving XGBoost models, and provides an XGBoost model implementation for prediction, pre and post processing. In addition, model lifecycle management functionalities like liveness handler, metrics handler etc. are supported.

To start the server locally for development needs, run the following command under this folder in your github repository. Also please ensure you have installed the kserve before.

make dev_install

Once XGBoost server is up and running, you can check for successful installation by running the following command

python3 -m xgbserver
usage: __main__.py [-h] [--http_port HTTP_PORT] [--grpc_port GRPC_PORT]
                   --model_dir MODEL_DIR [--model_name MODEL_NAME]
__main__.py: error: the following arguments are required: --model_dir

You can now point to your model file and use the server to load the model and test for prediction. Models can be on local filesystem, S3 compatible object storage, Azure Blob Storage, or Google Cloud Storage.

Install the development dependencies with:

make dev_install

To run tests:

make test

To run static type checks:

mypy --ignore-missing-imports xgbserver

An empty result will indicate success.

You can build and publish your own image for development needs. Please ensure that you modify the inferenceservice files for XGBoost in the api directory to point to your own image.

To build your own image, navigate up one directory level to the python directory and run:

docker build -t docker_user_name/xgbserver -f xgb.Dockerfile .

Sometimes you may want to build the XGBServer image with a different version of XGBoost, you can modify the version "xgboost == X.X.X" in setup.py and build the image with tag like docker_user_name/xgbserver:1.1.0.

To push your image to your dockerhub repo,

docker push docker_user_name/xgbserver:latest